I would like to thank Grégoire Lacaze and Aurélie Ceccaldi for inviting me to take part in the research seminar on reported speech and subjectivity they convened at Aix-Marseille Université, and for the very warm welcome they accorded me on my visit to Aix-en-Provence. I am also very grateful to E‑rea’s anonymous referees for their very careful reading and helpful suggestions.
1The field of what was once commonly referred to as “computer-mediated communication” – a term that sounds quaintly outdated now – has broadened out significantly since its early beginnings. The increasing availability, in large parts of the world, of Internet access and (often mobile) devices used for this access has given rise to considerable growth and diversification of the forms of online communication, well beyond the websites, message boards or e-mails of yesteryear. Fields such as communication studies, linguistics and stylistics face a challenge trying to study these forms seriously, adapting and updating existing frameworks or devising new ones where traditional models of discourse and conversation analysis may not always be fit for purpose. This is a challenge worth rising to, not just for its inherent intellectual interest, but also for at least two ethical imperatives. First, most researchers are teachers as well, and we owe it to our students to take an active research interest in forms of online behaviour they are often deeply engaged in. Second, in an age of unparalleled communicative possibilities, there appear to be similarly unparalleled risks, for instance in terms of deliberate misinformation and of deliberate stoking of mistrust among people of different backgrounds and beliefs.
2Against this background, this paper revisits the vast and oft-studied topic of speech and thought representation, and the expression of subjectivity therein, through the relatively narrow lens of a number of online manifestations of it. I will turn first to a number of uses of quotation, whether marked by quotation marks and/or by the use of the quotative be like, in Internet memes and tweets. Next I consider the broader use of the conversation frame (Pascual & Sandler 2016) in various iterations of fictive mini-dialogues in Twitter discourse. In a third section I offer a preliminary examination of the fascinating area of quote-tweeting, in which a Twitter user retweets a previous tweet and adds a comment, whether in textual, visual or hybrid form. The concluding section tries to recognize both what is “old” and what is “new” in these various online modes of using quotation or conversation templates, not, or not merely, to actually quote individual utterances or conversations, but to express attitudes and elicit responses from others to these attitudes.
3As just suggested, for the purposes of this paper, I will consider both Internet memes and Twitter discourse, as in both cases properties of online media exploit existing forms of speech and thought representation in medium-specific ways. Twitter as a social media platform allows users to share opinions in short textual messages, possibly expanding these into lengthier threads, and possibly enhancing their messages by adding emoji to their text and by appending pictures, animated pictures or video clips to their tweets; Twitter users can interact by responding to tweets, with replies appearing underneath the tweet being responded to, or by retweeting a tweet on their own profile page or ‘timeline’ either with a comment, which then appears above the quoted retweet, or without (see section 3 below).
4Internet memes have attracted increasing attention in recent years (e.g. Shifman 2014, Milner 2016), with studies across the fields of communication science, rhetoric and sociolinguistics (e.g. Davison 2012, Shifman 2012, 2013, Milner 2013, Jenkins 2014, Wiggins & Bowers 2015, Huntington 2016, Yus 2018) and in cognitive linguistics and construction grammar more generally (e.g. Dancygier & Vandelanotte 2017, Lou 2017, Zenner & Geeraerts 2018, Bülow, Merten & Johann 2018). Shifman defines an Internet meme as “a group of digital items sharing common characteristics of form, and/or stance, which were created with awareness of each other, and were circulated, imitated, and/or transformed via the Internet by many users” (2014 41). This definition can cover a broad variety of phenomena, including for instance spoofs of video content that has gone viral, but I will restrict attention here to image files including text, as will become clear below. Internet memes can be circulated on various social media platforms, including Facebook, Twitter and Reddit; sometimes signature phrases or constructional markers used in some of them catch on beyond the online world and make it into more traditional discourses, including newspaper or advertising language (see Dancygier & Vandelanotte 2017 573, 576-577, 590).
5This paper’s aims are to provide a qualitative analysis of uses of quotation and dialogue formats in Internet memes and Twitter discourse, based on an extensive and growing collection of examples of both memes and tweets. Given the considerable creativity exhibited in much online discourse, I make no claims to exhaustiveness, but endeavour to highlight and analyse forms and functions of Internet memes and tweets that are particularly relevant to how Internet users’ subjectivities are expressed through uses of speech and thought representation forms for purposes other than, or additional to, reports of actually occurring specific utterances or thoughts.
6Speech and thought representation has long been recognized as not necessarily involving the reproduction of an actual prior ‘original’ utterance or thought – leaving aside the question whether this ‘reproduction’ is verbatim, as is required for instance in academic, legal or parliamentary reporting contexts, or more approximative or periphrastic (see, for example, von Roncador 1980, 1988; Tannen 1989; Vandelanotte 2009 Chapter 4). This includes various kinds of modalized, future, hypothetical, typified or negated uses, in examples such as I might have said X (but I didn’t); in such a case, a psychologist would say X; when that happens, the government will say X; nobody ever said X; and so on. Generalizing across usage, we can state that in using speech or thought representation constructions, language users ‘simply’ want to represent something as speech or thought, without this necessarily actually being the case. If it isn’t, then other functions come into play pragmatically: generally speaking, we can say some content is being held up for consideration, to elicit responses, attach attitudes to them, and the like. Not infrequently across languages, typologists have been able to identify ‘non-quotative’ usage of direct speech constructions in particular which has grammaticalized to express other meanings, including mental and emotional states, desires, intentions, attempts, states of affairs, causation, reason, purpose and future; for an overview of such cases, see Pascual (2014 Ch. 4) and Pascual & Sanders (2016).
7A prime example of quotation being used in Internet discourse not to actually quote anyone, but, in this case, to hold up attitudes for scrutiny and mockery, is the said no one ever or said no X ever meme. The meme has been around for some time, and was analysed in Dancygier & Vandelanotte (2016 28-34) as a case of viewpoint reversal: the quoted utterance at first appears to be attributable, typically to the character or characters portrayed in the image contained within the Internet meme, but the addition of said no one ever (or more specific variants as in said no student ever) subverses that initial viewpoint, exposing it as one that is too far-fetched, incredible or irrelevant for anyone (or any student, etc.) to seriously entertain it. The apparent quote is thus shown to be unattributable: there is no possible speaker to whom the utterance could be attributed. Once you zoom out from the two clashing viewpoints – the apparent quote, and the said no one ever clause which reverses it – you obtain a more general understanding, for instance, in the case of Figure 1 below, that no one likes always being an empathetic listener and never having the favour returned. In cases which feature a common noun such as student, teacher, gamer, girlfriend, etc. instead of one, as in “I can’t wait for class to start” said no student ever (Figure 2), the overall effect is to comically confirm and perpetuate stereotypes about groups of people.
Figure 1. A “said no one ever” meme
Figure 2. A “said no X ever” meme
8Communication in general is, of course, interactive to at least some degree; this is clearest in face-to-face interaction, but even the writer of the most dense and abstruse niche novel, for instance, still wants to communicate something, and possibly wants some form of response, however indirect. Online forms of discourse do not have quite the same kind of immediate response potential, but they do not lag far behind, as responses to Internet memes being shared or tweets being sent can be near-instantaneous and, unlike typical adjacency pairs in conversation, they can keep coming long after the initial post was shared online. What is also clear is that Internet memes as in Figures 1-2, and other examples to follow below, in the main assume a reader-and-viewership that shares a common outlook: even if perhaps in reality there are students who are keen for class to start, or there are individuals who do not mind never being lent a sympathetic ear by people they are always ready to listen to, the memes tend to presume an intersubjective complicity in which members of the discourse community share beliefs and attitudes implicitly.
9A second series of ‘non-quotative’ memes centres around the quotative be like. Quotative uses of be like and related forms have been extensively studied from a language variationist perspective mainly, as epitomized in Buchstaller’s (2014) book-length study, which stresses among other things how the global spread of the variant is accompanied by local differences in the different places where it has been adopted, especially with respect to social factors regulating its use. It has also been an interesting testing ground to help rethink the structural analysis of how direct speech and thought constructions are put together syntagmatically (Vandelanotte 2009 Ch. 2, Vandelanotte & Davidse 2009), and to probe the suitability of the grammaticalization framework (e.g. Romaine and Lange 1991) to the emergence and spread of the be like quotative (Vandelanotte 2012).
10In its meme manifestation, the be like quotative is characterized by use of a plural subject noun phrase – most frequently the rather crude bitches and niggas but also many nouns less likely to give offence, such as dudes, moms, grandparents, kids, doctors, teachers, and the like (see Vandelanotte 2019 187-189). The plural underscores the stereotype-reinforcing nature of these memes, much like the said no X ever variant discussed above. The plural subject noun is followed by uninflected be like (a feature influenced by African American Vernacular English; see, for example, Cukor-Avila 2002) in the text appearing at the top of the meme, and an apparent quote in the text appearing at the bottom. In an example like Figure 3, the quote typifies behaviour of grandparents indulging their grandchildren’s cravings, and the clash between the text (“little snack”) and the image showing a gigantic Hershey’s chocolate tablet adds a humorous note, suggesting the attitude the meme user is expressing overall is one of friendly mockery. Examples often appear to have this humorous incongruity between the text and the image (see, e.g., Attardo 1994); one example for instance shows a woman on the look-out with binoculars accompanied by the text Bitches be like “I’m over him”; another has the text Moms be like “bring me my purse” over a picture of coathooks with a multitude of purses hanging from them.
Figure 3. A “be like” meme
11A less whimsical, more politically engaged use of the be like quotative to express attitudes online is found in Twitter uses exemplified in Figure 4. Again it is the base form be like rather than an inflected form which is used, and a generalizing, typifying characterization is being offered up of groups of things or people (for instance, “schools” in Figure 4, or “America”). The tweet combines ordinary typed text with emoji in a highly effective way, presenting schools as places where bullying, racism and homophobia are not cared about nor responded to (hence the emoji that expresses an “I can’t be bothered” shrug), but where comparatively minor offences such as chewing gum, wearing hats or wearing a shirt that shows your shoulders is frowned upon or even penalized (suggested by the use of a red-faced, angry-looking emoji). The tweet thus compares and contrasts stances schools purportedly take (not caring about important forms of discrimination, but enforcing petty rules about behaviour and dress code), with as ultimate aim to express disapproval over this perceived hypocrisy.
- 1 This tweet was no longer available online at the time of publication; more detailed source informat (...)
Figure 4. Using “be like” on Twitter1
12There is, in an example like Figure 4, no longer a single utterance-type ‘content’ similar to the traditional notion of a reported clause, as in I was like, how’s that been going for you?. Instead, a series of pithily evoked attitudes are being summarily depicted by combining key words or phrases with emoji. Figure 5 below returns to something closer to the meme format described previously (exemplified in Figure 3), but shares with Figure 4 the complexity of the content being (fictionally) quoted, which in this case is cartoon containing a further layer of speech exchange, between the knife and fork on the one hand, and the chopsticks on the other hand. The question “So… Which one of you is the fork?” might have seemed bizarre on its own, but the top line reading “Straight people be like” has opened up our frame of knowledge pertaining to straight and, by implication, non-straight relationships, allowing us to understand the scene, and the question asked within it, as an unusual but clever analogue for the situation gay people find themselves in when asked “which one of you is the man/woman?”. The overlay of facial features onto the kitchen utensils adds both a comic touch, and additional support for an animistic conception of these implements, in which it becomes possible to think of them as having a conversation. Ultimately, of course, knives and forks do not speak nor do they have faces, and the scenario is an unrealistic one. As Semino (2010) has shown, unrealistic scenarios can be used very effectively to teach us something about hard-to-grasp content, especially if some humour is thrown into the mix; what it teaches us here is empathy and a realization of how the perhaps well-meaning but misguided cliché question asked of gay couples is really not appropriate. The use of an apparently incongruent visual scene, serving as source domain, in combination with text, identifying at least partly the target domain, forms an interesting type of figurative language-and-image use which Lou (2017), in his study of when-memes, has referred to as multimodal simile, a notion which seems applicable to this example as well.
Figure 5. “Be like” introducing a complex, unrealistic dialogue in cartoon form
13Uses such as those illustrated in Figures 4 and 5 clearly stray some way from that in Figure 3: where the latter uses the be like quotative in the same way as it can be used in spoken interaction, the former two depict attitudes in ways which cannot easily be condensed into a “quoted clause”, not just because of the important role played by visual elements, but also because an internally highly complex content is being demonstrated in the part of the tweet or meme which follows the be like quotative. Nevertheless, there are still strong enough links with quotative be like, which after all has similative meaning origins, and which is often used to represent emotional responses and attitudes expressed gesturally and/or using interjections or non-lexicalized sounds (in examples such as And he was like [throw-away gesture]; she was like, wow! or I was like, eurgh). More general notions of demonstration (Clark & Gerrig 1990) or depiction (Clark 2016) can arguably cover both the more commonly studied uses of be like and these Internet uses to which it is put, and what is being demonstrated can potentially be complex and lengthy. One example posted on the Academia Obscura Facebook page showed the quotative Academic writing be like followed by a four-panel Calvin and Hobbes cartoon, in which Calvin is practising writing as incomprehensibly as possible, demonstrating a perceived abstruseness typical of academic writing. As a whole, then, the Facebook post lampoons academics’ self-importance and academic writing’s deliberate obfuscation.
14In addition to memes or tweets using existing direct speech constructions, such as the said no one ever and be like examples above, as a central part of their pre-set format, Internet memes using other formats may also, more incidentally, include quotes in the text they contain. One example is given in Figure 6, an example of the so-called “Scumbag Steve” set of memes in which the same picture – a fairly random picture which for reasons unknown came to represent “bad guys” in online discourse – is used to create memes typifying and calling out inconsiderate actions perpetrated by men who cannot be trusted (for a fuller treatment of this and related memes, see Dancygier & Vandelanotte 2017 577–586).
Figure 6. A “Scumbag Steve” meme using quotation
15The top and bottom text in this and related memes follow a neat division of labour in which the top text is understood as the “if P” or “when P” part of a predictive construction (Dancygier 1998), and the bottom text as the “then Q” part. Varying degrees of ellipsis tend to affect the text in this type of meme, as in an example like Uses shopping cart / leaves it in empty parking space, where both articles and subject noun phrases are left out – the subject of course being available visually instead of textually. The example in Figure 6 is particularly concise in its textual form, but proficient meme users who know the pattern that is being built on here will understand the top text as something “Scumbag Steve” would say in a typical situation, and the bottom text providing the dénouement in which the person buying drugs off Steve finds out they paid for more than they received. A full textual rendition of this could run something like When Steve sells you drugs and says “Yeah that’s like 3.5 grams” and you go home and check what you’ve bought, you’ll find it’s actually only 2 grams. All of this, of course, is part of a fictive scenario, making the quotation contained within the meme another kind of ‘non-actual’ quote, referred to by Pascual (2014) more generally as fictive interaction, where the frame of a conversation or interaction is used not to actually describe a conversation that took place, but to serve, in this case, as a way of exposing and ridiculing typical bad behaviour.
16A rather different kind of inherently ‘non-quotative’ usage is that of made-up or “fake” quotes, where something is deliberately mis-represented as having been uttered by someone, whether for purposes of misinformation or (as in Figure 7 below) for sarcastic commenting. In Figure 7, Twitter user Alex Hirsch sarcastically echoes a Twitter attack, in late July 2019, by President Trump on US Representative Elijah Cummings of Baltimore, who he accused of doing “NOTHING for his very poor, very dangerous and very badly run district”. The tweet uses both quotation marks and the convention of indicating the source of a quote at its end, following a long dash; in its typography (with its use of all caps for stress) and content the tweet closely echoes President Trump’s style, which, together with the source indication “cool prez [president] who loves America”, makes it abundantly clear who the mock quote is targeting. Clearly the effect of the fictive quote is not to mislead anyone, but to present an amplified picture of the president’s behaviour in order to ridicule and criticize it, an intention supported by the evocation of the aural effect of an (attention-seeking) airhorn.
17The example thus uses a number of clear conventions of quotation (the quotation marks, the dash followed by the source) to do something rather different: to express disapproval towards behaviour that is seen as both self-congratulatory and offensive towards others. While in other contexts fabricated quotes might be used maliciously, there is enough here to signal the fabrication, particularly in the “airhorn” sound effect being evoked and the clearly jocular formulation of the apparent quote’s source; in addition, the Twitter lay-out is such that the person inventing the fictive quote for comic purposes is very prominently identified by their name and Twitter account name immediately above the apparent quote. Figure 8 provides a similar type of case, but one which is different in important respects. Here the fabricated quote is not ‘faux-attributed’ to a specific high-profile individual like the US President, but is chosen to characterize the attitude or state of mind of people who are in favour of the UK exiting the EU in the most abrupt, ‘no deal’ type of way, come what may (or indeed, as the Prime Minister at the time of writing has it, “do or die”). In addition, the quote is juxtaposed to a brief video clip of a cyclist forcing his way through a level crossing barrier to cross, only to be forced to stop due to an oncoming train. The video clip is clearly not literally about Brexit, but is figuratively applied, as a case of multimodal simile (Lou 2017), to the frame of mind of people determined to impatiently press ahead with a course of action known to be dangerous and irresponsible.
Figure 8. Fabricated quote combining with video clip as a case of multimodal simile
[https://twitter.com/kidd__kong78/status/1177673256138096641]
18Where this section looked at the use of apparent quotation to not actually quote utterances, the next section broadens out onto the use of a type of dialogue format, again not to actually represent a dialogue or conversation, but to hold up often contrasting behaviours for scrutiny and criticism or ridicule.
19A popular meme circulating on various social media platforms revolves around the use of what ostensibly seems a mini-dialogue to in fact criticize contradictory attitudes in one and the same person, or to criticize other types of behaviour. The convention used in all of these examples is that in which participants in a dialogue or conversation are identified by a noun phrase followed by a colon; outside of online discourse this has long been found, for instance, in journalistic interview formats or sometimes in dialogue sequences in fiction. In the case where contrasting attitudes held by the same person at different times are being mocked or criticized, the pattern includes use of also, with self-criticism expressed in me: … also me:… examples as in Figure 9, and criticism of others (often public figures) expressed in examples such as Figure 10.
- 2 This tweet was no longer available online at the time of publication; more detailed source informat (...)
Figure 9. First person use of “X … also X…”2
20In Figure 9, the contrast in attitudes is expressed across verbal and visual modes, with the initial stance being expressed as a typified quote (an attitude ‘summarized’, so to speak, in the form of a quoted utterance), and the contradictory position being expressed in two images of cats showing behaviour which can be interpreted as clingy: the facial features in the first picture, and the attention-demanding physical posture in both, present types of feline behaviour which we can fairly easily understand as applying to human behaviour too (cf. Lou 2017 on when-memes, some of which rely on physically comparable features between animals and humans). In Figure 10, by contrast, the contradictory positions held by Conservative Party Member of Parliament James Cleverly are both expressed purely textually, using quotes which typify and summarize his political judgements at different times in order to criticize a perceived incoherence. Correlating with the difference in person (first person for self-directed, vs. third person for other-directed criticism), examples such as those in Figures 9-10 (and many others I’ve looked at) suggest that the “me” type of examples focus more on unimportant topics – referred to in Shifman’s (2012) study of YouTube memes as “whimsical content” – whereas the third person examples can be much more politically engaged and thus involve a serious critique which is meant to bite. Features of typification and generalization, commented on in the literature on speech and thought representation generally (e.g. Clark & Gerrig 1990 773, Fludernik 1993 Ch. 8), are still more strikingly present in an example as in Figure 11, where the speaker of the apparent quote is not an individual, but a vaguely defined collective, “partisan twitter”. It is precisely the feature of partisanship contained within the noun phrase identifying the apparent speaker which is being called out in the tweet, which points up the partisan way in which people evaluate the practice of digging up old tweets by politicians to criticize them – approval if it concerns a politician they dislike, disapproval otherwise.
Figure 11. Collective third person plural use of “X… also X…”
[https://twitter.com/robfordmancs/status/1098985164192055297]
21In including schematic slots for “politician I like” or “politician I dislike” in square brackets, for readers to fill in, the example continues the schematization begun in the source identification (“partisan twitter”), functioning much like a template. These slots could be said to suspend reference to any specific case in point, standing in for any more particular value people might have in mind; in this respect, they function a bit like items such as so and so and such and such in speech and thought representation contexts (Vandelanotte 2009 120-123).
22While the examples in Figures 9-11 feature the same apparent speaker or speaker collective being quoted at different times to criticize, in jest or more seriously, contradictory attitudes, other forms and uses of dialogic patterns involve interactions between different speakers. Figure 12 combines a recognizable type of message which many people will know from experience with online purchases (“your order has been shipped”) with a visual depiction of the tweeter’s response, using the “me” already shown in Figure 9. Not unlike what we also saw in that earlier example (Figure 9), the image used to represent the type of embodied behaviour in response to the order’s having been shipped in fact involves an animal (a dog) rather than a human. This again adds to the humour of the tweet, which gently holds up for mockery an online shopper’s intense wait by the front door so as not to miss the delivery man.
Figure 12. Typified quote followed by a response expressed through visual simile.
[https://twitter.com/garyfromteenmom/status/918324919276097537]
23A more frequent pattern, continuing the pattern shown in Figures 9-11, is that of fictive mini-dialogues, sometimes with only two turns as in the examples up to this point, sometimes with more. Often this use of the conversation frame serves to ridicule unnecessary, illogical, unasked for, or otherwise unhelpful pronouncements, as in the examples in Figures 13-14. Both examples suggest that people offer up a statement which was unprompted and likely not of much interest to anyone. In Figure 13, the suggestion is that author J.K. Rowling (of Harry Potter fame) seeks attention by revisiting published work and adding new interpretations to it, specifically in response to her hinting that the character Dumbledore was gay; the tweet amplifies and caricatures this with a rather extreme fictive quote being attributed to Rowling. Figure 14 illustrates a commonly used pattern where a number of ‘empty’ quotes (representing no one saying or asking anything at all) lead up to someone offering a response where none was wanted. In this case, the use of what one might call ‘internet troll capitalization’ for the noun ‘iNfLuEnCeR’ identifying the apparent speaker adds further force to the denigration of the category of people being criticized here – that of influencers who show and talk about products on social media platforms like Twitter or Instagram in exchange for freebies or indeed for pay.
Figure 13. Ridiculing a fictive unsolicited opinion offered by a specified individual
[https://twitter.com/bobvulfov/status/1107090381576040449]
Figure 14. Ridiculing a fictive unsolicited opinion offered by a type of person
[https://twitter.com/cxcope/status/1118243734716657667]
24In addition to text (or, as in Figure 14, the absence of text) being depicted following the source indication, tweets using this dialogue format can also use images (pictures, animated GIFs, video clips), as in Figure 15, or smileys and the emoji available directly within the Twitter platform, as in Figure 16, to represent attitudes or situations. In Figure 15, the Twitter user, who is an author who has published a first book, characterizes his response to a typical question about completing a second book by a using an animated GIF of a dancer along with four skeletons together dancing a dance macabre. Again, no single, actually occurring speech exchange is assumed to be reproduced here, but a typified situation is expressed with its accompanying emotional response. The visual means used in Figure 16 are, in a sense, simpler, using the available repertoire of emoji and the even more basic, typographically expressed forms of smiley faces (sad face and happy face) to express, wordlessly, how people’s moods can be lifted by contemplating a starry night sky. The use of the conversation frame here is further removed from any actual dialogue in a case like Figure 16, since words do not enter into it anywhere, the turns in the apparent ‘conversation’ here being either an attitude (sad or happy) or the mere presence of stars in the night sky. As an illustration of the very notion of fictive interaction, using apparent quotation to express subjective attitudes changing over time, however, it is very effective.
Figure 15. Fictive mini-dialogue incorporating an image
[https://twitter.com/caesura/status/1114566829610024960]
25In addition to the textual and combined textual-visual examples of tweets using the conversation frame to express and respond to attitudes, it is worth briefly mentioning that (as happens so often with Internet memes and other online communication formats) ‘meta-forms’ that show awareness of the format, as well as deliberately punning, exist. Figure 17 presents a simple but amusing case playing on the polysemy of the word colon and the expected use of the colon-as-punctuation-mark in the conversation frame tweet format. Figure 18 again involves a pun, in that it relies on the knowledge that the proper name William has a more popular counterpart or alternative in Bill, which is homonymous with the common noun bill; in achieving its humorous effect the example further relies on the use of apposition in the source designation me, an intellectual, highlighting the way the staged ‘speaker’ of the second line thinks rather highly of themselves.
Figure 18. Punning and attempted one-upmanship in the conversation frame tweet format
[https://twitter.com/Caolan_Walsh/status/1060925840148652033]
26The use of the conversation frame to structure attitudes and responses to them as mini-dialogues is pervasive on Twitter, and for many users of the platform has become second nature. It can use both text and visuals, whether images external to the platform or the limited repertoire of smileys and emoji built into it. Compared to the form I turn to in the next section – that of quote-tweeting – this format does not rely very strongly on special affordances of platform lay-out, but is simple enough to be effective on its own terms. Quote-tweeting, on the other hand, provides an extremely interesting case of using quotation for subjective commenting on expressed stances, as the platform-specific form this takes provides a built-in slot for commenting.
27Tweets can be responded to and interacted with in a variety of ways: you can “like” a tweet; you can reply to it by adding a comment (whether textual or visual or both), which then appears underneath the tweet; you can retweet it without comment (which places the retweeted tweet on your own ‘timeline’, i.e. the page which forms your Twitter ‘profile’); or you can retweet with comment. This option of retweeting with comment is also called quote-tweeting. This, of course, is “real” quoting, different from the practices surveyed in the previous sections, where something was being staged as an utterance or an interaction, without there being a specific underlying speech event with a unique speaker or set of interlocutors, in order to typify attitudes and depict responses. Here the Twitter platform offers a method of quoting an existing tweet alongside the quoting tweeter’s response to it. It is not that this is absent from “offline” speech and thought representing strategies; in English, for instance, it is possible to use an adverb modifying the verb used in a reporting clause, whether of direct (1) or indirect (2) speech or, in (3), thought:
(1) “I thought you said he was in England,” I stupidly said. “I said we believe he was,” Wesley said. (COCA corpus, Davies 2008–)
(2) As a result, the News of the World and News International wrongly maintained that these issues were confined to one reporter. (iWeb corpus, Davies 2018–)
(3) Everyone who foolishly thought the trophy was already in the hands of the engravers with specific instructions (that’s “Durant” with a T at the end and not a D) is now suddenly and very quickly backtracking a bit. (COCA corpus, Davies 2008–)
28The adverbs stupidly, wrongly and foolishly overlay the reported utterance or thought with the reporting speaker’s current appraisal (in example (1), pertaining to their own previous utterance). Other languages may show subtle constructional strategies to convey a meaning of mistaken belief, as shown by Spronck & Si (2019). What is particularly interesting about quote-tweeting, however, is the very flexible and creative way in which the quoting tweet can comment on the quoted tweet, revealing the quoting tweeter’s stance towards the latter, as the examples to follow will illustrate.
29Before considering some of the varieties of quote-tweeting, it is worth pointing out that the practice of quote-tweeting is not always felt to be neutral or appropriate, and considerable metadiscursive negotiation can attach to it, with Twitter users commenting on each other’s or their own ways of producing discourse (on metadiscourse in this reflexive sense, see e.g. Mauranen 1993, Ädel 2006). A common belief is that it is better to avoid quote-tweeting objectionable tweets from objectionable accounts, felt to be driving hate speech, as this draws further attention (and online traffic) to the views and people objected to; screenshots are often used to allow Twitter users to still respond to such content without the undesirable side-effects of amplifying the reach of someone whose views they object to. This has the additional potential benefit of preventing the quoted Twitter user from seeing their message is being interacted with (unless, of course, another Twitter user decides to tag the author of the quoted tweet in a comment, something sometimes objected to as ‘snitch-tagging’).3 In other cases, people explicitly explain why they decide to quote-tweet something (4), before going on to provide a factual correction of the quoted tweet (not shown here), or they question the quoting Twitter user’s motives in quote-tweeting instead of engaging directly through responses (5-7); in the case of (7), the offending Twitter user responded by saying “Because I want more people to see it”:
(4) Gonna QT [quote tweet] this because I’m seeing it shared a lot. [followed by a correction to the quoted tweet] [https://twitter.com/addadc/status/1174717067196555264]
(5) In the end to another extremely normal week, I return from the pub to find that Andrew Neil has spent the evening quote-tweeting a playground-style row I didn’t know I was having with him on the subject of whether white supremacy is a thing [followed by screenshots of multiple cases of journalist Andrew Neil quote-tweeting and responding to journalist Jonathan Lis] [https://twitter.com/jonlis1/status/1173009417387892736]
(6) You know what you’re doing. You’re trying to portray me as a bully by *retweeting* [in fact, retweeting with comment, i.e. quote-tweeting, LV] my replies rather than replying. If you really cared you’d stand up for what you preach. [https://twitter.com/samjmckee/status/1172796494841286656]
(7) Why did you quote tweet this instead of replying like a normal human being [https://twitter.com/joncstone/status/1170605057211011072]
30The tone of justification or requests for justification illustrated in these examples already suggests that where quote-tweeting is concerned, it is often serious, sometimes heated discussions that are at stake, rather than the whimsical, humorous content many of the examples in previous sections concerned. Where said no one ever, be like, me… also me… and many ‘mini-dialogue’ examples of online discourse often assume that the authors and readers of these messages share a common outlook, and enjoy poking fun at a range of attitudes, this is not the default discourse situation in the case of quote-tweeting, however amusing and creative some of the comments may be. While quote-tweeting can be, and is regularly, used to express support for the quoted tweet, by adding some concurring thought or appreciative phrase for instance, or used to express surprise about a situation, it is also, probably more frequently, used to express disagreement and criticism. For technical reasons, for a long time it was only possible to use text in the quote-tweet comment ‘slot’; this, as we will see below, led to some creative innovations in which visual responses were alluded to textually. More recently, in the course of 2019, the technical restriction was lifted, and quote-tweet comments can now include visuals. I will turn first to a number of textual types of comments and return briefly to a few visual examples at the end of this section.
31The first two examples include the kind of fictive dialogue frame we have encountered in the previous section in the quote-tweet comment slot. In Figure 19, the apparent ‘dialogue’ between “Monday” and “Friday” is used to suggest a further escalation in a series of political crises, with news of a referral of the Prime Minister for possible misconduct in public office breaking on the Friday of a week already rich in drama and upheaval in UK politics. The pretend quote “hold my copy of constitutional law” refers to the popular cliché “hold my beer”, used to announce that the speaker will enter the fray or do something even more outrageous than someone else has just done.
32Figure 20 is more complex in that the quote-tweet comment first makes a direct comment (“I’m roaring at this”) and clarifies an aspect of the tweet which it is quoting approvingly (“nobody even asked”), expressing hilarity at the statement in which a British Member of Parliament feels the need to respond, completely unprompted, to the controversy raised over old photos of Canadian Prime Minister Justin Trudeau in which the latter wears blackface. The quote-tweet comment subsequently repeats this point (“nobody even asked”) in fictive interaction form, introducing the fictive dialogue (of the “unsolicited opinion” type we saw in section 2) explicitly by means of the phrase It’s literally like.
Figure 20. Quote-tweet comment including “it’s literally like” and conversation frame
[https://twitter.com/dylanhm/status/1177520937601851392]
33Where the examples in Figures 19-20 use fictive dialogue formats in the quote-tweet comments, the examples in Figures 21-23 use both real and fictive individual quotes. In Figure 21, a real quote, in which then President of the European Council Donald Tusk advises the UK not to waste the time of an extension granted to the UK for its withdrawal from the EU, is used to provide implicit comment on a tweet describing a discussion on word etymology between the Leader and the Speaker of the House of Commons, clearly suggesting time is not being used wisely at all. Pitching the Tusk quote against the realities of an irrelevant debate in the House of Commons clearly reveals the quoting tweeter’s critical stance.
- 4 This account is sometimes locked, which is why the Twitter account and URL have been withheld, but (...)
Figure 21. Quote-tweet comment pitching genuine quote against the quoted tweet.4
34The example in Figure 22 uses the pretence of a quote to in fact paraphrase the quoted tweet so loosely as to reformulate it entirely according to the quoting Twitter user’s interpretation of it, namely as a purely attention-seeking message with no further content worth considering. Note that here, technically speaking, we do not have a case of quote tweeting in the strict sense, given that the quoting Twitter user incorporates a screenshot in their own tweet, rather than retweeting the original message with comment, for the kinds of reasons outlined above – not wanting to drive traffic to the Twitter account, and/or not wanting to engage directly with the author of the embedded tweet.
Figure 22. Apparent quote, in fact a deliberately very loose paraphrase, as quote-tweet comment
[https://twitter.com/StevePeers/status/1155078496529592320]
35The examples in Figures 23-25 are different from the previous ones in very closely echoing, in the quote-tweet comment, the structure of the quoted tweet, in order to express a dissociative attitude towards the latter. Echoing has long been analysed in Relevance Theory (e.g. Sperber & Wilson 1986, Iwata 2003) in terms of the combination of a metarepresentation with an attitude expressed towards this metarepresentation (cf. Vandelanotte 2010), and the characterization works particularly well in the context of echoic quote-tweeting practices. The example in Figure 23 shares with the previous two examples the use of a quote (here, as in Figure 22, an apparent rather than a genuine quote), in this case marked by quotative be like. The quoting tweet echoes the original, quirky tweet put out by the social media team at Virgin Trains in order to criticize the high ticket prices for which the company is held responsible, as well as hinting at its fear of (re-)nationalization.
36The echoic examples in Figures 24-25 do not involve (real or apparent) quotation in the quote-tweet comment slot itself, but of course still provide a comment on the quoted tweet by closely echoing its form with, essentially, sarcastic intent. In Figure 24, scientist and pro-EU campaigner Mike Galsworthy castigates Conservative Party Member of Parliament Michael Fabricant, known for his remarkable hairdo (hence the reference to a hairdresser), for rather brutally taking on a Twitter user who had suggested that Boris Johnson would not be able to get his way, and that Brexit, being impossible to deliver, would end the UK. The example in Figure 25 shows an impressively elaborate example of an echoic quote-tweet comment, in which the quoting Twitter user suggests the quoted tweet is wrong to use the results from a past UK General Election (2017 in this case) to argue about the likely electoral fortunes of the Labour Party in a much anticipated potential 2019 or 2020 election. In order to make the point that in an extremely volatile political climate, with unpredictable effects of up to four parties polling at 15 to 30% in a first past the post system, the quoting Twitter user Oz Katerji echoes the quoted tweet line by line (with the exception of the final line), using the 1832 election result and drawing the obviously absurd conclusion that the only party with a realistic chance of delivering a new referendum on the final Brexit question is the long defunct Whig Party.
37When Twitter users are content to comment merely with an emotional reaction to a quoted tweet, the technical restriction which until recently meant that no images could be used in a quote-tweet comment led to an innovation, illustrated in Figures 26-27, by which a textual description was offered of a corresponding visual artefact. In Figure 26, reference is made to a well-known animated GIF image showing a dog (with human features) surrounded by flames saying “this is fine”, based on a comic.5 The mention of “this-is-fine.gif”, adopting the file name format most computer users will be familiar with, is sufficient to metonymically evoke the corresponding image for those Internet users who interact on social media often enough to be familiar with this particular meme. The format used in Figure 26 is slightly different, as the text used to allow readers to call up the corresponding animated GIF – involving a character from the popular sitcom Gavin and Stacey saying “Oh my Christ!” – is not shown in the syntax of a file name as it would appear as listed, but rather as it could be read out in dictation.
Figure 26. Text-for-GIF as file format used as quote-tweet comment
[https://twitter.com/chrishanretty/status/1082242508858511360]
Figure 27. Text-for-GIF in dictation format used as quote-tweet comment
[https://twitter.com/JamieAli_/status/1154073424349794304]
38As a side note, it is worth mentioning that the text-for-GIF device can involve completely fictive descriptors of GIFs. For instance, in response to a tweet sharing an article by Professor Steve Peers entitled “The witch-hunt against an anti-gay professor is an own-goal for the left”, someone responded with a tweet reading “Karl Popper paradox of tolerance dot gif”;6 there is not actually (to the best of my knowledge) a widely shared GIF image relating to this theorem.
39A related type of creative device by which a different semiotic mode is called up metonymically by means of a textual description is that in which an apparent name of a sound file is used to evoke the corresponding sound; Figure 28 illustrates this with the descriptor “Sad_trombone.wav”, in mock sympathy with a tweet (since deleted) which it in fact ridiculed.
40In the course of 2019, Twitter rolled out technical improvements which mean that the comment accompanying a quoted tweet need no longer be restricted to text only, freeing the way to quote-tweet comments including visuals, as in the examples in Figures 29-30 below. In Figure 29, the quote-tweet comment uses a fictive conversation where the first turn is textual, but the second is given as a visual, itself incorporating the text “Well yes, but actually no” in answer to the question whether they (i.e. the Tory or Conservative Party) want to avoid a no-deal exit of the UK from the EU. The equivocation represented in this ‘yes but no’ answer already suggests criticism, and indeed the broader situation of discourse makes it clear that the quoting Twitter user, a member of the Scottish Greens, is criticizing the stance expressed in the quoted tweet, authored by a member of the Scottish Conservatives. In Figure 30, the quote-tweet comment does not rely on the conversation frame, but uses a multimodal simile (Lou 2017, Dancygier & Vandelanotte 2017) where the animated GIF image of someone repeatedly stepping on a garden rake is seen as providing a model for the interpretation of how the Labour Party has been groping around for a credible position on Brexit. This combination of text and image as a whole then functions as a dismissive comment on the interview fragment and its description posted by journalist Robert Peston, who was interviewing the Labour Shadow Chancellor.
- 7 Account and URL info withheld as this is from a smaller type of account; more detailed source infor (...)
Figure 29. Quote-tweet comment incorporating a static image as part of a fictive interaction.7
Figure 30. Quote-tweet comment featuring a multimodal simile with an animated GIF as its visual component
[https://twitter.com/matthewcarson/status/1154335826593099776]
41Across the examples analysed in this section, the quote-tweet comment types on display have been many and varied, including the use of apparent or real quotes, fictive dialogues and close (and sometimes elaborate) echoes of the quoted tweet, with the medium used in the comment being initially textual only, but also, in creative variations, text standing in for other media (image and sound), and finally, more recently, also incorporating images. The device of quote-tweeting and the way it is implemented technically and presented visually makes it an extremely versatile locus for the expression of the quoting Twitter user’s subjective responses to the quoted tweet and, one might say, ‘upgrades’ or ‘grammaticalizes’ a position only fairly minimally exploited in traditional offline quotation practices, as suggested in examples (1-3) above. In addition, as also discussed above, to the practice of quote-tweeting attach unspoken, but strongly felt rules, the breaking of which can lead to fierce metadiscursive debate, as in examples (5-7) above. For all these reasons, quote-tweeting seems to me a fascinating area for researchers interested in emerging and developing styles and uses of quotation to explore and investigate in greater depth.
42This paper has surveyed a range of practices involving quotation to highlight and comment on subjective attitudes in online discourse. Often the quotation device itself was fictive, in that no actual specific quoted utterance or thought preceded the quotation, which was instead being used to represent attitudes and stances and to respond to these, often critically and/or humorously, as in uses of said no one ever or be like memes, other uses of quotation marks in existing meme formats, uses of be like in tweets, and various uses of the conversation frame in tweets. In the case of quote-tweeting, while the quote-tweet comment itself may again contain examples of fictive or ‘non-quotative’ quotation, the quote-tweet act itself is of course a genuine act of quotation, but one which includes in its format a space for commenting open to considerable creative exploitation across text and image, and image-text combinations. As the examples surveyed have shown, not all uses of (non-)quotation to express and respond to attitudes, behaviours and beliefs pertain to light-hearted or whimsical topics; more politically engaged uses are particularly common in the case of quote-tweeting and in third person uses of the “X… also X…” fictive interaction format. In addition, ridicule or humour does not, of course, preclude more serious engagement and criticism being expressed simultaneously, as in the case of the fabricated Trump quote (Figure 7). In terms of the forms used – often in embedded or stacked combinations and sometimes using textual and visual modes in tandem – and the subjective functions performed, there is much worth exploring for students of (non-)quotation in social media discourse.